Objective: To assess the accuracy of MRI to predict remaining lymph node metastases in patients with complete pathological luminal response (ypT0) after neoadjuvant therapy.
Methods: Data from a national registry were used. 19 patients with histopathologically remaining lymph node metastases (ypT0N+) were identified. Another 19 patients without lymph node metastases (ypT0N0) were used as matched controls. Two radiologists blinded to all patient information evaluated staging and restaging MRI that was compared to histopathological findings of the resected specimen.
Results: The average size of the largest lymph node on restaging MRI was significantly larger (4.5 mm) in the ypT0N+ group than in the ypT0N0 group (2.6 mm) (p = 0.04). Presence of ypN+ was correctly predicted by MRI in 7 of 19 patients. In patients without lymph node metastases (ypT0N0), these were correctly classified by MRI in 16 of 19 patients. All patients who had MR-identified lymph nodes larger than 8 mm at restaging were ypTN+. The sensitivity, specificity, positive predictive value and negative for prediction of remaining lymph node metastasis with MRI were 37, 84, 70 and 57%.
Conclusion: In patients with ypT0 in rectal cancer after neoadjuvant treatment, remaining regional lymph node metastases cannot safely be predicted by restaging MRI alone using presently known criteria. Presence of a lymph node over 8 mm on restaging MRI strongly indicates yPN+. Advances in knowledge: This is one of the first studies on MRI lymph node assessment after chemo-radiotherapy (CRT) in luminal complete response.